A novel prediction method for outlet water temperature of converter valve based on F-BP network

Bo Peng,Xiaohui Liu,Hui Sun,Jinjin Ding,Kaipei Liu,Jing Wang,Sihan Zhou,Liang Qin
DOI: https://doi.org/10.1016/j.egyr.2023.05.142
IF: 5.2
2023-10-01
Energy Reports
Abstract:A converter valve is the core equipment of HVDC transmission system, whose operating temperature threshold is strictly constrained. This paper proposes a novel prediction method for outlet water temperature of converter valve based on F-BP network, which aims to accurately predict the outlet water temperature and assists the operation and maintenance personnel to take measures in time so that the temperature of the converter valve will not exceed its preset threshold when the operation condition has changed. Firstly, the principle and method of the construction of typical operation databases of converter valve is stated, including data standardization, the calculation of the optimal clustering category number and the final clustering process. Then, the steps of using the typical operation databases and BP neural network to make predictions are presented. Using MATLAB, we predicted the outlet water temperature of a converter valve in Chuxiong Converter Station with F-BP method and two other existing methods in comparison. The results indicate that the proposed approach’s prediction accuracy increases by 0.9141 °C and 0.9938 °C respectively compared with the simple BP neural network and linear regression, which contributes to the prediction application of the outlet water of a converter valve.
energy & fuels
What problem does this paper attempt to address?